The arithmetic is simple. Capacity in milliamp-hours divided by average draw in milliamps gives hours. A 2000 mAh cell powering an 80 mA load lasts 25 hours on paper.
runtime (h) = capacity (mAh) × usable fraction / current (mA)
In practice you will not get that. The derating factor in the calculator accounts for the gap, and 0.7 to 0.85 is a fair starting range.
Why the paper number is optimistic
- Cutoff voltage. Your circuit stops working before the cell is empty. A regulator that drops out at 3.0 V leaves real energy in a lithium cell that discharges to 2.5 V.
- Rated versus actual capacity. The printed figure is measured at a low, steady discharge rate. Pull harder and you get less, because internal resistance wastes some of it as heat.
- Self-discharge. Alkaline cells lose a few percent a year; NiMH can lose that much a month unless they are low-self-discharge types. For a device that sleeps for a year, this can dominate everything else.
- Temperature. Capacity falls sharply below freezing. A cell rated 2000 mAh at 20 °C may deliver 1200 mAh at −10 °C.
- Ageing. Lithium-ion cells lose roughly 20% of capacity over a few hundred full cycles.
Averaging a duty-cycled load
Most battery-powered designs are not constant loads. If a sensor node draws 120 mA for 200 ms every 30 seconds and 15 µA the rest of the time, the average is (0.2/30)×120 + (29.8/30)×0.015 ≈ 0.81 mA. That is the number to divide into the capacity. The sleep current usually matters more than the active current, which is why cutting sleep draw from 100 µA to 10 µA can multiply battery life several times over.